Harvested shoots were dried at 60°C, weighed, ground with a plant mill, and stored for further investigations. The P content in plant tissue was measured after dry ashing using the molybdovanadate method (Page et al. 1982). Plant P uptake was calculated by multiplying the P content of the shoots and shoot biomass.

The soil samples were air-dried and sieved (2 mm) before analysis. Soil pH was measured in 0.01 M CaCl2 using a 1:2.5 soil-to-solution ratio. For characterization of soil P pools, different methods were used. The method described by Van der Paauw (1971) was used to determine water-extractable P (Pw) with a soil-to-water ratio of 1:25. The P concentrations in the extracts were measured by the phosphomolybdate blue method via flow-injection analysis. The content of double-lactate-soluble P (Pdl) (photometric method) was quantified according to Blume et al. (2000). By means of the ammonium oxalate method (Schwertmann 1964) the extractable amount of P (Pox) allows the estimation of the amount of inorganic P being adsorbed on amor­phous iron and aluminium oxides in the soil. Pox and the oxalate-soluble aluminium and iron contents (Alox, Feox) in soil were extracted and their concentrations were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES; JY 238, Jobin Yvon, France). With use of these data, the P-sorption capacity [PSC (mmol kg-1) = (Alox + Feox)/2] and the degree of P saturation [DPS (%) = Pox/ PSC x 100] could be calculated according to Lookman et al. (1995) and Schoumans (2000). Total P was analysed after aqua regia dissolution in a microwave oven (Mars Xpress, CEM, Kamp-Lintfort, Germany) followed by ICP-OES.

Furthermore, the sequential P fractionation method developed by Hedley et al. (1982) was used. Different P fractions of decreasing bioavailability (resin-P, NaHCO3-P, NaOH-P, H2SO4-P) were extracted step by step by using stronger extracting agents. The remaining P in the soil sample after the extraction steps is considered as residual P. The residual P content was determined by subtracting the amount of extracted P from the total P content [residual P = total P — (resin-P + NaHCO3-P + NaOH-P + H2SO4-P)] as described by Schlichting et al. (2002). Total P was determined by aqua regia digestion in a microwave oven.

2.2.2 Statistics

Soil and plant data corresponding to four spatial replications were subjected to two- and one-factorial analysis of variance (general linear model). The results are reported as main effects and interactions. The means of soil and plant parameters were compared by the Duncan test. Significances were determined at p < 0.05. Significantly different means were indicated by using different characters. The statistical analysis was carried out using SPSS 15.0.